Inversion of well logs into facies accounting for spatial dependencies and convolution effects

Lindberg, David Volent; Rimstad, Eivind; Omre, Henning

doi:10.1016/j.petrol.2015.09.027

Cited by 22 publications

(8 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Eidsvik, Mukerji, and Switzer (2004), a hidden Markov model is applied to estimate the facies profile and the facies transition probabilities from well log. Similarly in Lindberg, Rimstad, and Omre (2015), the EM method is combined with hidden Markov models. However, we point out that these methods can be difficult to apply in a 3D real reservoir characterization study because of the large number of required spatial parameters.…”

Section: Discussionmentioning

confidence: 99%

Statistical facies classification from multiple seismic attributes: comparison between Bayesian classification and expectation–maximization method and application in petrophysical inversion

Grana

Lang

2016

Geophysical Prospecting

View full text Add to dashboard Cite

We present here a comparison between two statistical methods for facies classifications: Bayesian classification and expectation–maximization method. The classification can be performed using multiple seismic attributes and can be extended from well logs to three‐dimensional volumes. In this work, we propose, for both methods, a sensitivity study to investigate the impact of the choice of seismic attributes used to condition the classification. In the second part, we integrate the facies classification in a Bayesian inversion setting for the estimation of continuous rock properties, such as porosity and lithological fractions, from the same set of seismic attributes. The advantage of the expectation–maximization method is that this algorithm does not require a training dataset, which is instead required in a traditional Bayesian classifier and still provides similar results. We show the application, comparison, and analysis of these methods in a real case study in the North Sea, where eight sedimentological facies have been defined. The facies classification is computed at the well location and compared with the sedimentological profile and then extended to the 3D reservoir model using up to 14 seismic attributes.

show abstract

Section: Discussionmentioning

confidence: 99%

Statistical facies classification from multiple seismic attributes: comparison between Bayesian classification and expectation–maximization method and application in petrophysical inversion

Grana

Lang

2016

Geophysical Prospecting

View full text Add to dashboard Cite

show abstract

“…The traditional DNN classifiers, which take the measurement set of each depth as a spatial elementwise independent event, always ignore the shoulder bed effect of the well logging data (Lindberg et al ., 2015; Tian et al ., 2019a). Also, the DNN architecture introduces too many model parameters into the classification system.…”

Section: Methodsmentioning

confidence: 99%

“…Limited by the resolution of the logging tools, the well logging measurement at each depth step is a weighted summation of the logging responses at the corresponding and adjacent depths. That is, the vertical spatial couplings exist between the neighbouring measurements of the same logging curve (Lindberg et al, 2015;Tian et al, 2019a;Tian et al, 2020). However, this critical characteristic is always ignored by most of the proposed models which take the samples as independent events.…”

Section: Introductionmentioning

confidence: 99%

Deep learning assisted well log inversion for fracture identification

Miao

Li²,

et al. 2020

Geophysical Prospecting

View full text Add to dashboard Cite

Manual fracture identification methods based on cores and image logging pseudopictures are limited by the expense and the amount of data. In this paper, we propose an integrated workflow, which takes the fracture identification as an end-to-end project, to combine the boundary detection and the deep learning classification to recognize fractured zones with accurate locations and reasonable thickness. We first apply the discrete wavelet transform algorithm and a boundary detection method named changing point detection to enhance the fracture sensibility of acoustic logs and segment the whole logging interval into non-overlapping subsections by estimating boundaries. The deep neural network based auto-encoders and the convolutional neural network classifier are then implemented to extract the hidden information from logs and categorize the subsections as the fractured or non-fractured zones. To validate the feasibility of this workflow, we apply it to the logging data from a real well. Compare with the benchmarks provided by the support vector machine , random forest and Adaboost model, the one-dimensional well profile predicted by the proposed changing point detection-deep learning classifier is more consistent with the manual identification result.

show abstract

“…Reference [17] provides a recent overview of theory and several applications. References [18][19][20] demonstrate applications of HMM to borehole data of petrophysical variables collected to understand rock type alternation styles in the context of exploring for oil and gas resources. There are of course several alternatives to HMM for predicting the hardness variations of rocks from borehole data.…”

Section: Hidden Markov Modelsmentioning

confidence: 99%

Rock Classification Using Multivariate Analysis of Measurement While Drilling Data: Towards a Better Sampling Strategy

2018

View full text Add to dashboard Cite

Measurement while drilling (MWD) data are gathered during drilling operations and can provide information about the strength of the rock penetrated by the boreholes. In this paper MWD data from a marble open-pit operation in northern Norway are studied. The rock types are represented by discrete classes, and the data is then modeled by a hidden Markov model (HMM). Results of using different MWD data variables are studied and presented. These results are compared and co-interpreted with optical televiewer (OTV) images, magnetic susceptibility and spectral gamma values collected in the borehole using down-the-hole sensors. A model with penetration rate, rotation pressure and dampening pressure data show a good visual correlation with OTV image for the studied boreholes. The marble class is characterized by medium penetration rate and medium rotation pressure, whereas the intrusions are characterized by low penetration rate and medium to high rotation pressure. The fractured marble is characterized by high penetration rate, high rotation and low dampening pressure. Future research will use the presented results to develop a heterogeneity index, develop an MWD-based 3D-geology model and an improved sampling strategy and investigate how to implement this in the mine planning process and reconciliation.

show abstract

Inversion of well logs into facies accounting for spatial dependencies and convolution effects

Cited by 22 publications

References 18 publications

Statistical facies classification from multiple seismic attributes: comparison between Bayesian classification and expectation–maximization method and application in petrophysical inversion

Statistical facies classification from multiple seismic attributes: comparison between Bayesian classification and expectation–maximization method and application in petrophysical inversion

Deep learning assisted well log inversion for fracture identification

Rock Classification Using Multivariate Analysis of Measurement While Drilling Data: Towards a Better Sampling Strategy

Contact Info

Product

Resources

About